Testing modified (Horndeski) gravity by combining intrinsic galaxyalignments with cosmic shear

TL;DR

This paper investigates how combining intrinsic galaxy alignments with cosmic shear measurements can improve constraints on modified gravity theories within the Horndeski class, highlighting the potential and limitations of this approach.

Contribution

It demonstrates that intrinsic alignments can enhance constraints on Horndeski gravity parameters, but their effectiveness depends on the calibration of alignment models and their correlation with gravity parameters.

Findings

Intrinsic alignments can improve Horndeski gravity constraints by 10% with simulation-based alignment parameters.

Self-calibration of intrinsic alignments shifts sensitivity primarily to alignment strength, reducing benefits for gravity tests.

The methodology can be extended to study other modifications of gravity and related cosmological parameters.

Abstract

We study the impact of modified gravity of the Horndeski class, on intrinsic shape correlations in cosmic shear surveys. As intrinsic shape correlations (IAs) are caused by tidal gravitational fields acting on galaxies as a collection of massive non-relativistic test particles, they are only sensitive to the gravitational potential, which forms in conjunction with the curvature perturbation. In contrast, the cosmic shear signal probes the sum of these two, i.e. both Bardeen-potentials. Combining these probes therefore constitutes a test of gravity, derived from a single measurement. Focusing on linear scales and alignments of elliptical galaxies, we study the impact on inference of the braiding $\hat{\alpha}_B$ and the time evolution of the Planck mass $\hat{\alpha}_M$ by treating IAs as a genuine signal contributing to the overall ellipticity correlation. We find that for \textsc{Euclid}, IAs can help to improve constraints on modified gravity of the Horndeski-class by 10 per cent if the alignment parameter needed for the linear alignment model is provided by simulations. If, however, the IA needs to be self calibrated, all of the sensitivity is put into the inference of the alignment strength $D$ since there is a very strong correlation with the evolution of the Planck mass. Thus diminishing the benefit of IA for probing modified gravitational theories. While the present paper shows results mainly for modified gravity parameters, similar deductions can be drawn for the investigation of anisotropic stresses, parameterised modifications to the Poisson-equation, the phenomenology of gravitational slip and to breaking degeneracies in a standard cosmology.